Cell cultures, consisting of cells growing suspended in a growth media, or on the surface of suspended particles, in solution are produced within bioreactors with careful control of a number of parameters. These bioreactors may be capable of processing large quantities of cell culture solution. For example, large-scale bioreactors can have capacities from 1-20,000 litres, or even up to 50,000 litres.
Within the bioreactor it is important to carefully control the environment to which the cells are exposed. Subtle changes in the environment can have major effects on the physiology of the cells and the amount of the target product (product titre), for example a recombinant protein, that is produced by each cell. This in turn has a major impact on the economics of the production process. The parameters that must be controlled include the concentrations of oxygen and carbon dioxide available to the cells (dissolved oxygen and CO2), pH, temperature, and specific nutrient levels such as the concentration of glucose. Additionally the physical environment is critical; particularly important components including the form of the gas distribution e.g. bubble size and overall gas flow. Finally, the mixing of the liquid and cells is critical having an impact on the homogeneity within the reactor and hence the local environmental variation to which cells within a bioreactor are exposed. Such issues become significant in very large bioreactors.
The major challenge facing companies manufacturing products in bioreactor systems is the optimisation of the conditions within a bioreactor for the production of a particular product. Optimisation of conditions for a particular cell line producing a particular product can easily have magnitude level effects on the yield of the product, this in turn having a massive impact on the economics of production. Addressing this issue is not simple; there are many parameters to be controlled and the optimal approach may involve variations in these conditions over time. However, it is impractical to explore the impact of varying a range of parameters due to the lack of availability of equipment and the huge costs of operation. The actual costs of one run of a 21 bioreactor can be over $2000. At larger scales the cost rapidly becomes prohibitive. Such issues prevent the application of modern statistical based experiment approaches to resolving the impact of multiple parameter variation typically referred to as DOE (Design of Experiment), such approaches typically requiring tens of bioreactor experiments to have value.
The opportunity for such work to have value has increased over recent years as regulatory authorities have introduced initiatives in which variations within a production run do not necessarily mean the automatic failure of a batch IF the impact of such variations in control parameters has previously been explored. This is impossible without small-scale highly parallel models of bioreactors but essential for manufacturers to remain competitive.
A further issue faced by bioreactors is the difficulty of selecting cell lines early in development that are robust and productive in a stirred bioreactor environment. Clearly, where high tens to hundreds of cell lines need to be screened, existing bioreactor systems are impractical.
A number of small-scale approach bioreactors have been tried, e.g. shaken multiwell plates and flasks, but these lack the ability to faithfully reproduce the conditions found in stirred, gassed systems with closed loop control of culture parameters. To date, small-scale experiment runs are generally carried out in individual bioreactors, of 1 to 10 litre capacity, containing cell cultures in solution. These are processed under careful, monitored control for a period of about two weeks. During that period, the input parameters discussed above may be varied between the individual bioreactors, with the contents of the respective bioreactors being monitored so as to determine which set of parameters achieves optimum, desired results. That set of parameters can then be used in order to scale-up the process to full production scale; the objective being to maximise cell production or cell viability, to improve production efficiency and/or to increase product titre yield.
Control of the culture parameters is required from three perspectives: i) the maintenance of a parameter at a defined set-point, within control limits, for a given time; ii) the controlled, planned variation of that parameter over time; and finally iii) the consistency and reproducibility of that parameter from bioreactor to bioreactor and run to run. Once such control is achieved, parameters can be varied and the impact of the variation on productivity determined.
The cell culture solution within the bioreactor is stirred in order to ensure homogeneity. The rate of stirring can have a major impact on the productivity of the culture through the impact of the physical environment of the cells, for example shear, on the viability and productive life of the cells. Additionally, the stirring rate has a direct effect on mixing and therefore the efficiency of mass transfer of gasses from the input stream of bubbles into the liquid phase where it is available to the cells. The balance between stir rates and their potential negative effects and the benefits of good mixing and gas transfer must be established for a particular culture. At manufacturing scale, energy inputs to the reactor additionally become an important economic consideration.
In many existing small-scale systems, the contents of the bioreactor vessels are not stirred, but are instead agitated by shaking. Whereas this simplifies the system, the vessels not requiring individual stirrers, it does not produce accurate simulation of production scale conditions, in which the contents are stirred; shaking does not replicate the shear forces induced in the vessel contents by stirring. Additionally, gas transfer in shaken vessels is primarily through surface aeration rather than bubbles fed into the base of the system, altering the dynamics of the gas transfer and the physical environment.
Where stirrers are provided, each is independently driven from a drive source. It is time-consuming for the operator to connect and disconnect each stirrer to the associated drive source, as is required between experiment runs
There are two key aspects to the gas control within bioreactors: that of CO2 and that of O2.
The dissolved oxygen level in the bioreactor must be maintained at a set level to ensure a consistent availability to the cells such that metabolism is not limited. Typical maintenance levels vary between 15 and 50% of the maximum dissolved oxygen level achieved by air saturation. Approaches to achievement of this vary between users, some preferring to use lower input concentrations and higher flow rates, others higher input concentrations and lower flow rates. Control of the input flow rate is critical as it affects the stripping of other gases such as CO2 from the culture media.
The concentration of CO2 that the cells are exposed to can have significant effects on metabolism. Control of CO2 is additionally used to control pH in combination with bicarbonate based buffer systems in the media. Bubbles are also a key source of damage to cells and hence control of the total gas inflow rate is an important factor in maintaining cell viability.
The pH level within the bioreactor should remain within predetermined bounds, which can vary as the cell culture develops. Generally this is achieved by a combination of a bicarbonate based buffer system within the liquid media, combined with the maintenance of a specific level of dissolved CO2. However, above a certain cell density the production of lactic acid by the cells can overwhelm the buffering capability of the media and the pH is maintained within the desired limits by the addition of doses of alkali solutions to combat the increasing acidity. The addition of alkali in bioreactors is controlled as part of a feedback loop including a pH sensor.
Temperature is an important parameter within bioreactors. The temperature used within bioreactors culturing mammalian cells does not vary widely due to the origins of the cells in animals exhibiting control of body temperature. However, some minor variations are used during the period of culture, to effect shifts in metabolism biasing the cell physiology towards production of the recombinant protein rather than cell multiplication for example.
Generally, a heater is controlled in order to increase or decrease the amount of supplied heat. In some systems the culture growth and energy inputs into stirring generate excess heat, so cooling and heat dissipation systems are required.
A range of nutrient feeds may be dispensed into the reactor. Typically these include media feeds which supply additional amino acids and carbon sources to replace those used in cell growth. Multiple different feeds may be added to a bioreactor on different schedules, often including pure carbon sources such as glucose. Generally, such feeds are added in response to the measurement of parameter levels within the bioreactor.
Monitoring of various parameters within the bioreactor is key to their control. Some parameters are controlled through closed loop sensing and response systems, others through sampling and off-line analysis due to the lack of appropriate on-line monitoring systems.
On-line monitoring systems are of two types: invasive and non-invasive. Invasive sensors rely on a probe carrying a sensor being inserted into the vessel and having direct contact with the culture solution. Generally, such systems are reusable and must be cleaned and calibrated between uses. Such monitors contribute to the complexity of setting up bioreactors through the requirement for disassembly from the reactor vessel for cleaning and sterilisation and the requirement for aseptic assembly. Some probes can be sterilised with the vessel but do require removal of residues and cleaning. Non-invasive sensor systems are now available in which a non-disposable sensing component has no contact with the culture, therefore does not require cleaning, sterilisation and accompanying validation of those processes.
One on-line monitoring method is to include disposable sensor spots in the vessel for remote interrogation. For example, a pH and/or dissolved oxygen sensor patches attached to the inside of the vessel can be interrogated externally; the spot is illuminated by a light source and a light detector detects emitted fluorescence, the characteristics and dynamics of which are indicative of the pH or dissolved oxygen levels within the vessel. Other techniques are available, including measurement of light diffraction and reflectance of near-infrared light. In this context, light is defined as encompassing any emission within the electromagnetic spectrum, not just visible wavelengths.
In current small-scale systems where sensing is done by non-invasive interrogation, individual vessels each have respective associated sensing apparatus, such as a light source and associated detector at each vessel location. This is expensive and also problematic in that the individual sensors must each be calibrated to ensure that the monitoring of each vessel is consistent.
The monitoring of the vessel contents may be achieved by invasive methods in which a small sample portion of the cell culture solution is removed for on-line or off-line analysis, for example via sampling port or by aspirating a sample of the solution with a pipette for dispensing for example into the sample cup of an analytical system. Likewise, a sample portion of the gases in the headspace within the vessel may be extracted for analysis in, for example, a gas analyser. That extraction may be done by a probe inserted into the headspace, or via a gas outlet port and associated conduit.
Where the liquid samples and/or the headspace gases are analysed via extraction through an outlet port in each vessel, it is time-consuming and often manually complex for an operator to connect and disconnect the fluid conduits to the respective outlet ports, a primary risk being the contamination of the device to be inserted. Such problems add to the general complexity and cost of conducting multiple bioreactor experiments.
In summary, there is a range of challenges in the development and optimisation of bioreactor based manufacturing processes, including: i) general costs of operation of current systems, even that of small scale systems being prohibitive due to complexity of set-up, labour, capital cost, equipment availability within facilities infrastructure required (steam generation) and high costs of media components per unit volume; ii) lack of directly applicable small-scale systems to model larger bioreactors; and iii) a lack of trained personnel driving the requirement for improved throughput per trained employee.
Accordingly, it is an object of the invention to improve the efficiency of the turnaround between experiment runs in microscale bioreactor systems, reduce labour requirement and increase throughput in laboratories.
It is another object of the invention to provide a more cost-effective solution to the monitoring of multiple individual vessels in a microscale bioreactor system.